Académique Documents
Professionnel Documents
Culture Documents
2011
ABSTRACT
Several grindability tests were developed over the years to design grinding circuits or optimize existing
operations. Each test has its own strengths and weaknesses and it is imperative to select the proper test
procedure(s) to meet project deliverables and minimize the risk of a project. Pilot plant testing of large bulk
samples historically constituted the traditional approach for AG or SAG design, but was gradually replaced
by small-scale tests. Nowadays, conducting grindability tests requiring only a few kilos of material on several
samples, is a more typical approach to grinding mill design. This paper summarises the requirements and
deliverables of various bench-scale test procedures, their strengths and weaknesses, and cases where AG/SAG
pilot testing should still be performed.
KEYWORDS
SUMMARY
INTRODUCTION
The resistance of ore samples to
breakage (or hardness) is measured
through grindability testing. Several
grindability tests have been developed
over the years for different applications
and each test has its own strengths and
weaknesses. Grindability testing is a
compromise between test cost and its
deliverable(s). Because a large fraction
of the cost component is driven by the
sampling requirement, tests that can
be performed on small drill cores offer
a significant cost advantage over those
that require large diameter drill cores
and substantial weight. On the other
hand, the test deliverables are generally
superior for tests requiring more weight.
Overviews of grindability testing
methodology and the compromise
between test sample requirement and
deliverables were discussed in previous
SAG conferences by Mosher and Bigg
(2001 & 2002) and McKen and Williams
(2006), introducing the now updated list
presented in Table 1.
GRINDABILITY TEST
MILL
DIA.
TOP SIZE
(M)
CLOSING
SIZE
SAMPLE
REQUESTED1
SAMPLE
CONSUMED2
TYPE
STEADYSTATE
DATABASE
(Y/N)
(Y/N)
(MM)
(CORE)3 (MM)
(KG)
(KG)
76.2
PQ/HQ
N/A
25
10
Single Particle N
Media Competency
1.83
165
N/A
750
300
Batch
MacPherson
Autogenous
0.46
32
NQ
1.18
175
100
Continuous
JK Drop-weight
N/A
63
PQ/HQ
N/A
75
25
Single Particle N
SMC Test
N/A
31.5
Any
N/A
20
Single Particle N
JK Rotary Breakage
Test
0.45
53
HQ
N/A
75
15
Single Particle N
SAGDesign
0.49
38.1
NQ
1.7
10
Batch
SPI
0.305
38.1
NQ
1.7
10
Batch
AG Pilot Plant
1.75
200
Various
>50,000
>50,000
Continuous
Lab-scale HPGR
0.25
12.7
BQ
3.35
400
360
Locked-cycle
SPT
N/A
19.1
BQ
3.35
10
Locked-cycle
0.9
50
Various
>2,000
>2,000
Continuous
0.305
12.7
Any
1.18
15
10
Locked-cycle
0.305
3.35
Any
0.149
10
Locked-cycle
Mod Bond
0.305
3.35
Any
N/A
1.2
Batch
1
2
Approximate weight consumed in the test for typical ores (S.G. = 2.8g/cm3).
Minimum whole core size required for a complete test. Partial results can sometimes be obtained with smaller cores.
The recommended top size for an SMC test is 31.5mm, but the test can be performed on smaller rocks or drill core, requiring smaller
weights.
4
Rotor Diameter.
Includes 250kg for a series of 7 batch tests to determine the optimal operating conditions and 150kg for a locked-cycle test.
0.01
0.1
1
10
Size K80 (mm)
100
1,000
Table 2 Approximate Range of Size Reduction in Comminution Tests, Devices and Processes
GRINDABILITY TESTS
The following is an updated review of
the principal grindability tests that are
currently commercially available for ore
characterization and their application
to circuit design. It is presented as
a reference guide and the reader is
encouraged to consult the references
that are more specific to each individual
test. All of these tests are supported by
fairly large databases (McKen, Verret, &
Williams, 2006).
MACPHERSON AUTOGENOUS
GRINDABILITY TEST
The MacPherson autogenous
grindability test was developed by
Arthur MacPherson (MacPherson
& Turner, 1978; McKen & Chiasson,
2006), as a continuous test performed
in a 46cm (18) semi-autogenous mill,
with an 8% ball charge. A draft fan
supplies the airflow required to remove
the ground material from the mill, and a
collection system recovers the ground
material from the air stream. This
includes a vertical classifier, a cyclone
and a dust collector (baghouse). The
cyclone underflow is classified on
a 14 mesh screen with the oversize
returning to the mill. The mill is fed
from a feed hopper by a Syntron feeder
actuated automatically by a Milltronics
control system. This control system
continuously regulates the feed rate
by maintaining a pre-set sound level
with a microphone located below the
mill shell, controlling the mill level to
25% charge by volume. The circulating
load is controlled to 5% by adjusting
the airflow through the mill. The test
requires material with a top size greater
than 32mm (1-1/4), and sufficient
weight to operate until all of the steadystate conditions are met, and for a
minimum of six hours. This can normally
be achieved with less than 100kg, but
typically, a 175kg sample is requested
to allow for soft and/or dense ores.
The test is run continuously, similar to
a small pilot plant, for a minimum of six
hours and until steady state is achieved.
At test completion, all of the products
are submitted for particle size analysis,
and the mill charge is dumped and
observed. The charge is submitted to a
particle size analysis as well as size-bysize S.G. determinations. This allows
the evaluation of any preferential coarse
build-up or particle density concentration
in the mill charge. The mill power draw,
throughput and product size distribution
are used to compute a specific energy
input and the MacPherson Autogenous
Work Index (AWI). Although the
importance of achieving steady-state
in a grinding test is widely accepted
(Bond tests), the MacPherson test
remains the only small-scale AG/SAG
mill test that offers this option. Steadystate is especially important in
AG/SAG mills where a harder
component can build up over time and
affect the production negatively.
SMC TEST
The SMC Test was developed by
Steve Morrell (2004). It is an abbreviated
drop-weight test, which can be
performed at low cost on small rocks
or drill cores (cores can be cut into
cylinders using a diamond saw). The test
is performed similarly to the standard
drop-weight test procedure, except that
a single size fraction is tested. The test
can be performed at various rock sizes,
the minimum acceptable top size being
16mm. The recommended particle
size is -31.5/+26.5mm, which requires
preparing about 30kg of samples, with
only 5kg actually tested, and all the
products and unused material can be
re-used for metallurgical testing. Testing
of smaller rocks or drill core requires
significantly smaller weight. A bulk
sample, or essentially any size of drill
core, is adequate for the test. The test
generates the Drop-weight Index (DWI)
expressed in kWh/m3, as well as the
A and the b parameters, but it does
not generate the crusher parameters,
which must be obtained through a full
drop-weight test. The test also provides
an estimated value of the ta, as well
as the Mia, Mic and Mih parameters,
more recently developed by Morrell
(2009). Normally, the main ore zones/
types in the deposit are submitted to
the full drop-weight test procedure and
the SMC Test is used to measure the
SAGDESIGN TEST
The SAGDesign test was developed
by the SAGDesign Consulting Group
(Starkey, Hindstrom & Nadasdy, 2006)
and consists of a batch grinding test
conducted in a 0.488m diameter SAG
mill. About 10kg of drill core is required
for testing. The feed is prepared to
80% passing 19mm and ground to 80%
passing 1.7mm. The SAGDesign product
is then crushed to 100% passing
3.35mm and used for a Bond ball mill
grindability test. The direct output of
the SAGDesign test is the number of
revolutions required to achieve 80%
passing 1.7mm. The SAGDesign and
Bond ball mill test results are used
by Starkey & Associates to design
commercial mills.
CONCLUSION
Simple tests requiring low sample
weights can be used for AG/SAG
variability testing and geometallurgical
mapping of an ore deposit, but they
have to compromise on the deliverables.
More sophisticated tests can provide a
more accurate and complete picture of
ore grindability, but they require more
material, so they can only be performed
on a minimum of samples.
It is highly desirable to submit all the
major ore types or alterations from a
deposit to a detailed characterisation
which covers the entire size range of
comminution. The Bond low-energy
impact test can be used to measure the
hardness at coarse size (up to 76.2mm),
while the variation of ore hardness by
size can be measured in the range 13.2
to 63mm using the JK drop-weight
test. The DWT results may be used
to extrapolate potential problems at
coarser size or to calibrate the tests
that can only be performed at finer size,
such as the SMC. The MacPherson
autogenous grindability test, which is
a steady-state test and cost-effective
pilot plant alternative, should also be
performed, because it will show if a hard
component of the ore, with same or
different specific gravity than the feed,
builds up over time, and if it causes
throughput problems. The Bond rod
mill and ball mill, as well as the Bond
abrasion tests should also be performed
on the main ore types to measure the
hardness at finer sizes and to evaluate
the ore abrasivity. Variability in the
deposit should be addressed through
a proper program. SPI and/or SMC
tests can both be used to test hardness
variability at AG mill size, while the
Bond ball mill grindability (or Mod Bond)
test remains the most appropriate way
to measure hardness at ball mill size.
The number of samples to be tested
will largely depend on the project size
and economics, as well as the level of
acceptable risk. High throughput/low
grade projects will require the highest
amount of testing. The combination
of methodologies will increase the
confidence in the design.
REFERENCES
SGS Minerals Services (2006).
Comminution Handbook.
Barratt, D.J., Matthews, B.D. & deMull,
T. (1996). Projection of AG/SAG mill
sizes, mill speeds, ball charges and
throughput variation from Bond work
indices. In A.L. Mular, D.J. Barratt
& D.A. Knight (Eds.), Proceedings
international autogenous and
semiautogenous grinding technology,
1996, Vol. 2, (pp. 541-558), Vancouver,
BC: Department of Mining Engineering,
University of British Columbia.
Barratt, D.J. & Doll, A. (2008). Testwork
programs that deliver multiple data sets
of comminution parameters for use in
mine planning and project engineering.
In R. Kuyvenhoven, C. Gomez, &
A. Casali (Eds.), Procemin2008, V
international mineral processing seminar,
2008, (pp. 81-95), Santiago, Chile:
GECAMIN Ltd.
Bond, F.C. (1947). Crushing tests by
pressure and impact. Trans AIME, Vol.
169, 58-66: The American Institute of
Mining, Metallurgical and Petroleum
Engineers.
Bond, F.C. (1960). Crushing and
grinding calculations. British Chemical
Engineering, Vol. 6, (Rev. January 1961
by A/C Pub. 07R9235B).
Bond, F.C. (1963). Metal wear in
crushing and grinding. Allis-Chalmers
Publication 07P1701.
Bulled, D. & Husain, K. (2008). The
development of a small-scale test to
determine work index for high pressure
grinding rolls. In Ian Orford (Ed.),
Proceedings 40th annual meeting of the
Canadian mineral processors, 2008, (pp.
145-164); Ottawa, ON: CIM
Burford, B. D. & Niva, E. (2008).
Comparing energy efficiency in grinding
mills. In Ian Orford (Ed.), Metallurgical
plant design and operating strategies,
2008, (pp. 45-64); Perth, Australia:
AusIMM
CONTACT INFORMATION
Email us at minerals@sgs.com
www.sgs.com/mining